
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Assuming 'data' is the dictionary containing the historical data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300,
             260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390,
             360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430,
             480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430,
                   430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480,
                   430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180,
                   290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490,
            480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790,
            780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640,
            640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490,
                   490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750,
                   590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640,
                   670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Convert 'Month' column to datetime and set it as index
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Split the data into training (75%) and validation (25%) sets
train_size = int(0.75 * len(df))
train_data = df.iloc[:train_size]
validation_data = df.iloc[train_size:]

# Set the color palette and style for better visualization
sns.set(style="whitegrid")

# Plotting the training and testing data using seaborn for better visualization
for compound in df.columns:
    plt.figure(figsize=(12, 6))
    sns.lineplot(data=train_data[compound], label='Training Data', color='blue', linewidth=2.5)
    sns.lineplot(data=validation_data[compound], label='Testing Data', color='red', linewidth=2.5)
    plt.title(f'Training and Testing Data for {compound}', fontsize=16)
    plt.xlabel('Month', fontsize=14)
    plt.ylabel('Price', fontsize=14)
    plt.legend(title='Data Type', title_fontsize='13', fontsize='12')
    plt.show()


import pandas as pd

# Define the data
data = {
    'Polymer Compounds': ['HDPE film', 'LDPE film', 'PET', 'PP homo-polymer fiber'],
    'Delivery Period (In days)': ['3-30 days', '3-30 days', '3-30 days', '3-30 days']
}

# Create DataFrame
df = pd.DataFrame(data)

# Print the DataFrame
print(df)

       Polymer Compounds Delivery Period (In days)
0              HDPE film                 3-30 days
1              LDPE film                 3-30 days
2                    PET                 3-30 days
3  PP homo-polymer fiber                 3-30 days


pip install graphviz

Collecting graphviz
  Downloading graphviz-0.20.3-py3-none-any.whl.metadata (12 kB)
Downloading graphviz-0.20.3-py3-none-any.whl (47 kB)
   ---------------------------------------- 0.0/47.1 kB ? eta -:--:--
   ---------------------------------- ----- 41.0/47.1 kB 991.0 kB/s eta 0:00:01
   ---------------------------------------- 47.1/47.1 kB 1.2 MB/s eta 0:00:00
Installing collected packages: graphviz
Successfully installed graphviz-0.20.3
Note: you may need to restart the kernel to use updated packages.


import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Assuming 'data' is the dictionary containing the historical data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300,
             260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390,
             360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430,
             480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430,
                   430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480,
                   430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180,
                   290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490,
            480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790,
            780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640,
            640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490,
                   490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750,
                   590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640,
                   670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Convert 'Month' column to datetime and set it as index
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)


df.describe()


df.head()


df.tail()

df


import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.metrics import mean_squared_error, mean_absolute_error

# Set the aesthetic parameters in one step
sns.set(style="whitegrid")

# Define the data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300, 260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390, 360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430, 480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430, 430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480, 430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180, 290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490, 480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790, 780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640, 640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490, 490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750, 590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640, 670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Create DataFrame from the data dictionary
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Interest rate for forward contracts
interest_rate = 0.0244  # 2.44%

# Delivery period for forward contracts
delivery_period = (3, 30)  # 3-30 days

# Define parameters and initialize variables for simulations
num_scenarios = 10000  # Number of scenarios for Monte Carlo simulation
n_months_future = 60  # Simulate future prices for the next 5 years (12 months * 5 years)
dt = 1 / 252  # Assuming daily data; adjust as needed

# Function to perform Monte Carlo simulation with specified parameters
def monte_carlo_simulation(s0, mu, sigma, n, num_scenarios, dt):
    prices = np.zeros((num_scenarios, n))
    for i in range(num_scenarios):
        price = np.zeros(n)
        price[0] = s0
        for t in range(1, n):
            simulated_return = (mu - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * np.random.normal()
            next_price = price[t - 1] * np.exp(simulated_return)
            if next_price <= 0:
                prices[i, t:] = np.nan
                break
            price[t] = next_price
        prices[i, :t+1] = price[:t+1]
    return prices

# Adjust the start date for simulation to April 2024
start_date_future = pd.Timestamp('2024-04-01')

# Initialize simulated prices dictionary
simulated_prices_future = {}

# Loop over each polymer compound and perform simulation
for material in df.columns:
    # Simulate future prices with specified parameters starting from April 2024
    s0_future = df[material].iloc[-1]
    mu = 0.003  # Drift (mu) of 0.3%
    sigma = 0.27124  # Volatility (sigma) of 27.12%
    simulated_prices_future[material] = monte_carlo_simulation(s0_future, mu, sigma, n_months_future, num_scenarios, dt)

    # Create a DataFrame for plotting
    dates = pd.date_range(start=start_date_future, periods=n_months_future, freq='M')
    scenarios = np.arange(num_scenarios)
    multi_index = pd.MultiIndex.from_product([dates, scenarios], names=['Date', 'Scenario'])
    plot_data = pd.DataFrame(simulated_prices_future[material].T.flatten(), index=multi_index, columns=[material])
    plot_data = plot_data.reset_index()

    # Plot the distribution of future prices for each material until Mar-2029
    plt.figure(figsize=(10, 6))
    sns.lineplot(data=plot_data, x='Date', y=material, hue='Scenario', palette='viridis', legend=None, alpha=0.2)
    plt.xlabel('Month/Year')
    plt.ylabel('Price')
    plt.title(f'Simulated Future Prices for {material} until Mar-2029')
    plt.ylim(0, None)  # Set y-axis limit to start from zero
    plt.show()

    # Calculate and print expected future prices at different time steps for each material until Mar-2029
    expected_prices_future = simulated_prices_future[material].mean(axis=0)
    print(f"\nExpected Future Prices at Different Time Steps for {material} until Mar-2029:")
    for month, price in zip(dates, expected_prices_future):
        print(f"{month.strftime('%B %Y')}: €{price:.2f}")

# Calculate and print MSE, MAE, RMSE, and MAPE for the overall model
def calculate_errors(actual, predicted):
    mse = mean_squared_error(actual, predicted)
    mae = mean_absolute_error(actual, predicted)
    rmse = np.sqrt(mse)
    mape = np.mean(np.abs((actual - predicted) / actual)) * 100
    return mse, mae, rmse, mape

# Assuming we have actual future prices for comparison (for demonstration purposes, using the last known prices)
actual_prices = df.iloc[-1].values

# Calculate errors for each material
for material in df.columns:
    predicted_prices = simulated_prices_future[material][:, -1]  # Using the last simulated prices
    mse, mae, rmse, mape = calculate_errors(np.full(predicted_prices.shape, actual_prices[df.columns.get_loc(material)]), predicted_prices)
    print(f"\nError Metrics for {material}:")
    print(f"Mean Squared Error (MSE): {mse:.2f}")
    print(f"Mean Absolute Error (MAE): {mae:.2f}")
    print(f"Root Mean Squared Error (RMSE): {rmse:.2f}")
    print(f"Mean Absolute Percentage Error (MAPE): {mape:.2f}%")

Expected Future Prices at Different Time Steps for HDPE film until Mar-2029:
April 2024: €1280.00
May 2024: €1279.86
June 2024: €1279.97
July 2024: €1279.99
August 2024: €1280.15
September 2024: €1279.79
October 2024: €1279.71
November 2024: €1279.63
December 2024: €1279.73
January 2025: €1280.15
February 2025: €1280.30
March 2025: €1280.40
April 2025: €1280.42
May 2025: €1280.28
June 2025: €1280.49
July 2025: €1280.63
August 2025: €1280.72
September 2025: €1280.44
October 2025: €1280.70
November 2025: €1280.57
December 2025: €1280.73
January 2026: €1280.67
February 2026: €1280.58
March 2026: €1280.26
April 2026: €1280.21
May 2026: €1279.99
June 2026: €1280.26
July 2026: €1280.37
August 2026: €1280.26
September 2026: €1280.60
October 2026: €1280.55
November 2026: €1280.77
December 2026: €1280.78
January 2027: €1281.09
February 2027: €1281.08
March 2027: €1281.05
April 2027: €1281.04
May 2027: €1281.27
June 2027: €1280.97
July 2027: €1281.14
August 2027: €1281.28
September 2027: €1280.91
October 2027: €1280.39
November 2027: €1280.66
December 2027: €1280.64
January 2028: €1280.70
February 2028: €1281.01
March 2028: €1281.02
April 2028: €1281.11
May 2028: €1281.55
June 2028: €1281.12
July 2028: €1280.67
August 2028: €1280.80
September 2028: €1280.65
October 2028: €1280.74
November 2028: €1280.61
December 2028: €1280.51
January 2029: €1280.39
February 2029: €1280.82
March 2029: €1280.84

Expected Future Prices at Different Time Steps for LDPE film until Mar-2029:
April 2024: €1060.00
May 2024: €1059.79
June 2024: €1059.94
July 2024: €1059.85
August 2024: €1059.75
September 2024: €1059.57
October 2024: €1059.39
November 2024: €1059.59
December 2024: €1059.73
January 2025: €1059.64
February 2025: €1059.03
March 2025: €1059.23
April 2025: €1059.42
May 2025: €1059.39
June 2025: €1059.09
July 2025: €1059.16
August 2025: €1059.32
September 2025: €1059.48
October 2025: €1059.46
November 2025: €1059.54
December 2025: €1059.09
January 2026: €1059.03
February 2026: €1059.06
March 2026: €1058.67
April 2026: €1058.70
May 2026: €1058.84
June 2026: €1058.67
July 2026: €1058.49
August 2026: €1058.47
September 2026: €1058.40
October 2026: €1058.47
November 2026: €1058.27
December 2026: €1058.05
January 2027: €1057.84
February 2027: €1058.06
March 2027: €1057.99
April 2027: €1057.91
May 2027: €1058.11
June 2027: €1058.47
July 2027: €1058.76
August 2027: €1058.83
September 2027: €1058.54
October 2027: €1058.70
November 2027: €1058.42
December 2027: €1058.43
January 2028: €1058.35
February 2028: €1058.52
March 2028: €1058.68
April 2028: €1058.80
May 2028: €1058.90
June 2028: €1058.78
July 2028: €1058.74
August 2028: €1059.00
September 2028: €1059.18
October 2028: €1059.38
November 2028: €1059.47
December 2028: €1059.63
January 2029: €1059.64
February 2029: €1059.60
March 2029: €1059.81

Expected Future Prices at Different Time Steps for PET until Mar-2029:
April 2024: €400.00
May 2024: €399.89
June 2024: €399.98
July 2024: €399.97
August 2024: €399.90
September 2024: €399.94
October 2024: €399.94
November 2024: €399.98
December 2024: €399.89
January 2025: €400.02
February 2025: €399.98
March 2025: €400.11
April 2025: €400.14
May 2025: €400.19
June 2025: €400.20
July 2025: €400.13
August 2025: €400.17
September 2025: €400.16
October 2025: €400.03
November 2025: €400.06
December 2025: €400.02
January 2026: €400.12
February 2026: €400.13
March 2026: €400.29
April 2026: €400.36
May 2026: €400.27
June 2026: €400.31
July 2026: €400.28
August 2026: €400.34
September 2026: €400.33
October 2026: €400.27
November 2026: €400.28
December 2026: €400.21
January 2027: €400.25
February 2027: €400.21
March 2027: €400.26
April 2027: €400.30
May 2027: €400.36
June 2027: €400.38
July 2027: €400.27
August 2027: €400.29
September 2027: €400.25
October 2027: €400.28
November 2027: €400.20
December 2027: €400.22
January 2028: €400.28
February 2028: €400.25
March 2028: €400.25
April 2028: €400.25
May 2028: €400.21
June 2028: €400.15
July 2028: €400.23
August 2028: €400.14
September 2028: €400.08
October 2028: €400.10
November 2028: €400.20
December 2028: €400.12
January 2029: €400.09
February 2029: €400.10
March 2029: €400.21

Expected Future Prices at Different Time Steps for PP homo-polymer fiber until Mar-2029:
April 2024: €1250.00
May 2024: €1249.82
June 2024: €1250.20
July 2024: €1250.09
August 2024: €1250.21
September 2024: €1250.49
October 2024: €1250.74
November 2024: €1250.95
December 2024: €1250.82
January 2025: €1250.89
February 2025: €1250.79
March 2025: €1251.14
April 2025: €1251.02
May 2025: €1251.16
June 2025: €1251.15
July 2025: €1251.56
August 2025: €1251.65
September 2025: €1251.83
October 2025: €1251.72
November 2025: €1251.53
December 2025: €1251.63
January 2026: €1251.84
February 2026: €1251.56
March 2026: €1251.50
April 2026: €1251.78
May 2026: €1251.71
June 2026: €1251.58
July 2026: €1251.65
August 2026: €1251.75
September 2026: €1251.84
October 2026: €1251.76
November 2026: €1251.77
December 2026: €1251.70
January 2027: €1251.87
February 2027: €1251.77
March 2027: €1251.85
April 2027: €1252.21
May 2027: €1252.14
June 2027: €1252.47
July 2027: €1252.26
August 2027: €1252.46
September 2027: €1252.44
October 2027: €1252.60
November 2027: €1252.27
December 2027: €1252.16
January 2028: €1252.20
February 2028: €1252.13
March 2028: €1252.56
April 2028: €1252.68
May 2028: €1252.46
June 2028: €1251.95
July 2028: €1252.23
August 2028: €1252.05
September 2028: €1251.86
October 2028: €1252.11
November 2028: €1251.93
December 2028: €1252.23
January 2029: €1252.26
February 2029: €1252.10
March 2029: €1252.27

Expected Future Prices at Different Time Steps for rPET until Mar-2029:
April 2024: €410.00
May 2024: €409.89
June 2024: €409.84
July 2024: €409.75
August 2024: €409.85
September 2024: €409.81
October 2024: €409.70
November 2024: €409.82
December 2024: €409.82
January 2025: €409.86
February 2025: €409.96
March 2025: €409.83
April 2025: €409.95
May 2025: €410.00
June 2025: €409.93
July 2025: €409.99
August 2025: €410.05
September 2025: €410.05
October 2025: €410.13
November 2025: €410.00
December 2025: €409.97
January 2026: €409.97
February 2026: €410.15
March 2026: €410.20
April 2026: €410.29
May 2026: €410.31
June 2026: €410.36
July 2026: €410.43
August 2026: €410.46
September 2026: €410.47
October 2026: €410.50
November 2026: €410.50
December 2026: €410.62
January 2027: €410.63
February 2027: €410.62
March 2027: €410.61
April 2027: €410.69
May 2027: €410.60
June 2027: €410.65
July 2027: €410.74
August 2027: €410.77
September 2027: €410.89
October 2027: €410.98
November 2027: €411.03
December 2027: €410.92
January 2028: €410.90
February 2028: €410.78
March 2028: €410.72
April 2028: €410.88
May 2028: €410.98
June 2028: €410.99
July 2028: €411.00
August 2028: €410.98
September 2028: €411.15
October 2028: €411.20
November 2028: €411.23
December 2028: €411.29
January 2029: €411.30
February 2029: €411.34
March 2029: €411.30

Expected Future Prices at Different Time Steps for rLDPE Film until Mar-2029:
April 2024: €470.00
May 2024: €469.92
June 2024: €469.80
July 2024: €469.78
August 2024: €469.91
September 2024: €470.04
October 2024: €470.01
November 2024: €470.02
December 2024: €470.11
January 2025: €470.00
February 2025: €469.90
March 2025: €469.90
April 2025: €469.84
May 2025: €469.73
June 2025: €469.62
July 2025: €469.61
August 2025: €469.58
September 2025: €469.58
October 2025: €469.61
November 2025: €469.61
December 2025: €469.34
January 2026: €469.31
February 2026: €469.20
March 2026: €469.24
April 2026: €469.14
May 2026: €469.35
June 2026: €469.21
July 2026: €469.31
August 2026: €469.22
September 2026: €469.29
October 2026: €469.33
November 2026: €469.27
December 2026: €469.34
January 2027: €469.31
February 2027: €469.23
March 2027: €469.23
April 2027: €469.31
May 2027: €469.39
June 2027: €469.54
July 2027: €469.50
August 2027: €469.40
September 2027: €469.43
October 2027: €469.38
November 2027: €469.40
December 2027: €469.51
January 2028: €469.50
February 2028: €469.46
March 2028: €469.34
April 2028: €469.42
May 2028: €469.42
June 2028: €469.28
July 2028: €469.23
August 2028: €469.11
September 2028: €469.08
October 2028: €469.16
November 2028: €469.25
December 2028: €469.24
January 2029: €469.20
February 2029: €469.19
March 2029: €469.36

Expected Future Prices at Different Time Steps for rPP until Mar-2029:
April 2024: €730.00
May 2024: €730.19
June 2024: €730.32
July 2024: €730.51
August 2024: €730.56
September 2024: €730.48
October 2024: €730.47
November 2024: €730.55
December 2024: €730.48
January 2025: €730.38
February 2025: €730.53
March 2025: €730.75
April 2025: €730.73
May 2025: €730.72
June 2025: €730.76
July 2025: €730.88
August 2025: €730.99
September 2025: €731.13
October 2025: €731.18
November 2025: €731.13
December 2025: €731.09
January 2026: €731.17
February 2026: €731.34
March 2026: €731.20
April 2026: €731.25
May 2026: €731.13
June 2026: €731.15
July 2026: €731.13
August 2026: €730.92
September 2026: €731.04
October 2026: €731.13
November 2026: €731.11
December 2026: €731.07
January 2027: €731.08
February 2027: €731.08
March 2027: €730.91
April 2027: €731.00
May 2027: €731.04
June 2027: €731.13
July 2027: €731.49
August 2027: €731.33
September 2027: €731.37
October 2027: €731.14
November 2027: €731.28
December 2027: €731.31
January 2028: €731.30
February 2028: €731.23
March 2028: €731.11
April 2028: €731.01
May 2028: €731.21
June 2028: €731.07
July 2028: €731.02
August 2028: €730.77
September 2028: €730.95
October 2028: €730.84
November 2028: €730.88
December 2028: €730.84
January 2029: €730.94
February 2029: €730.78
March 2029: €730.73

Expected Future Prices at Different Time Steps for rHDPE Film until Mar-2029:
April 2024: €520.00
May 2024: €520.02
June 2024: €520.02
July 2024: €520.06
August 2024: €520.08
September 2024: €520.06
October 2024: €520.18
November 2024: €520.20
December 2024: €520.18
January 2025: €520.12
February 2025: €520.03
March 2025: €520.06
April 2025: €519.91
May 2025: €520.05
June 2025: €520.18
July 2025: €520.18
August 2025: €520.17
September 2025: €520.13
October 2025: €520.09
November 2025: €520.00
December 2025: €519.99
January 2026: €519.96
February 2026: €519.99
March 2026: €520.09
April 2026: €520.05
May 2026: €519.99
June 2026: €520.02
July 2026: €519.96
August 2026: €520.09
September 2026: €520.07
October 2026: €520.03
November 2026: €520.01
December 2026: €519.96
January 2027: €519.85
February 2027: €519.55
March 2027: €519.42
April 2027: €519.45
May 2027: €519.46
June 2027: €519.36
July 2027: €519.27
August 2027: €519.39
September 2027: €519.39
October 2027: €519.41
November 2027: €519.46
December 2027: €519.49
January 2028: €519.58
February 2028: €519.54
March 2028: €519.45
April 2028: €519.42
May 2028: €519.45
June 2028: €519.44
July 2028: €519.42
August 2028: €519.45
September 2028: €519.56
October 2028: €519.58
November 2028: €519.63
December 2028: €519.56
January 2029: €519.52
February 2029: €519.53
March 2029: €519.42

Error Metrics for HDPE film:
Mean Squared Error (MSE): 28737.85
Mean Absolute Error (MAE): 134.94
Root Mean Squared Error (RMSE): 169.52
Mean Absolute Percentage Error (MAPE): 10.54%

Error Metrics for LDPE film:
Mean Squared Error (MSE): 19234.39
Mean Absolute Error (MAE): 110.07
Root Mean Squared Error (RMSE): 138.69
Mean Absolute Percentage Error (MAPE): 10.38%

Error Metrics for PET:
Mean Squared Error (MSE): 2871.20
Mean Absolute Error (MAE): 42.60
Root Mean Squared Error (RMSE): 53.58
Mean Absolute Percentage Error (MAPE): 10.65%

Error Metrics for PP homo-polymer fiber:
Mean Squared Error (MSE): 27153.81
Mean Absolute Error (MAE): 130.86
Root Mean Squared Error (RMSE): 164.78
Mean Absolute Percentage Error (MAPE): 10.47%

Error Metrics for rPET:
Mean Squared Error (MSE): 2878.34
Mean Absolute Error (MAE): 42.55
Root Mean Squared Error (RMSE): 53.65
Mean Absolute Percentage Error (MAPE): 10.38%

Error Metrics for rLDPE Film:
Mean Squared Error (MSE): 3885.92
Mean Absolute Error (MAE): 49.54
Root Mean Squared Error (RMSE): 62.34
Mean Absolute Percentage Error (MAPE): 10.54%

Error Metrics for rPP:
Mean Squared Error (MSE): 9369.66
Mean Absolute Error (MAE): 77.01
Root Mean Squared Error (RMSE): 96.80
Mean Absolute Percentage Error (MAPE): 10.55%

Error Metrics for rHDPE Film:
Mean Squared Error (MSE): 4656.94
Mean Absolute Error (MAE): 54.18
Root Mean Squared Error (RMSE): 68.24
Mean Absolute Percentage Error (MAPE): 10.42%


import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from scipy.stats import norm

# Set the aesthetic parameters
sns.set(style="whitegrid")

# Define the data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
}

# Create DataFrame from the data dictionary
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')  # Corrected date format
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Interest rate for forward contracts
interest_rate = 0.0244  # 2.44%

# Delivery period for forward contracts
delivery_period = (3, 30)  # 3-30 days

# Define parameters and initialize variables for simulations
num_scenarios = 10000  # Number of scenarios for Monte Carlo simulation
n_months_future = 60  # Simulate future prices for the next 5 years (12 months * 5 years)
dt = 1 / 252  # Assuming daily data; adjust as needed

# Function to perform Monte Carlo simulation with specified parameters
def monte_carlo_simulation(s0, mu, sigma, n, num_scenarios, dt):
    prices = np.zeros((num_scenarios, n))
    for i in range(num_scenarios):
        price = np.zeros(n)
        price[0] = s0
        for t in range(1, n):
            simulated_return = (mu - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * np.random.normal()
            next_price = price[t - 1] * np.exp(simulated_return)
            if next_price <= 0:
                prices[i, t:] = np.nan
                break
            price[t] = next_price
        prices[i, :t+1] = price[:t+1]
    return prices

# Adjust the start date for simulation to April 2024
start_date_future = pd.Timestamp('2024-04-01')

# Initialize simulated prices dictionary
simulated_prices_future = {}

# Loop over each polymer compound and perform simulation
for material in ['HDPE film', 'LDPE film', 'PET', 'PP homo-polymer fiber']:
    # Simulate future prices with specified parameters starting from April 2024
    s0_future = df[material].iloc[-1]
    mu = 0.003  # Drift (mu) of 0.3%
    sigma = 0.27124  # Volatility (sigma) of 27.12%
    simulated_prices = monte_carlo_simulation(s0_future, mu, sigma, n_months_future, num_scenarios, dt)
    simulated_prices_future[material] = simulated_prices

    # Calculate quantiles for each month
    quantiles = np.percentile(simulated_prices, [25, 50, 75], axis=0)

    # Plot quantile curves
    plt.figure(figsize=(10, 6))
    months = pd.date_range(start=start_date_future, periods=n_months_future, freq='M')
    plt.plot(months, quantiles[0], color='blue', label='25th Percentile')
    plt.plot(months, quantiles[1], color='green', label='50th Percentile (Median)')
    plt.plot(months, quantiles[2], color='red', label='75th Percentile')
    plt.title(f'Quantile Curves for {material}')
    plt.xlabel('Month')
    plt.ylabel('Price')
    plt.legend()
    plt.show()


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error, mean_absolute_error

# Set the aesthetic parameters in one step
sns.set(style="whitegrid")

# Define the data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300, 260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390, 360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430, 480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430, 430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480, 430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180, 290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490, 480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790, 780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640, 640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490, 490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750, 590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640, 670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Create DataFrame from the data dictionary
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Function to create lag features
def create_lag_features(df, compound_name, n_lags):
    lag_df = pd.DataFrame()
    for lag in range(1, n_lags + 1):
        lag_df[f'lag_{lag}'] = df[compound_name].shift(lag)
    return lag_df.dropna()

# Function to forecast prices using XGBoost
def forecast_prices_xgb(compound_name, forecast_period=60, n_lags=12):
    compound_df = df[[compound_name]].copy()
    n_lags = min(n_lags, len(compound_df) - 1)
    lag_df = create_lag_features(compound_df, compound_name, n_lags)
    y = compound_df[compound_name].iloc[n_lags:]
    X_train, X_test, y_train, y_test = train_test_split(lag_df, y, test_size=0.25, random_state=42)
    
    # Initialize and train the XGBoost model
    reg = xgb.XGBRegressor(n_estimators=1000, learning_rate=0.01, max_depth=3, objective='reg:squarederror', booster='gbtree')
    reg.fit(X_train, y_train, eval_set=[(X_train, y_train), (X_test, y_test)], verbose=False)
    
    # Forecast future prices
    future_dates = pd.date_range(start=df.index[-1] + pd.DateOffset(months=1), periods=forecast_period, freq='M')
    future_df = pd.DataFrame(index=future_dates, columns=[compound_name])
    last_known_data = lag_df.iloc[-1:].values.reshape(1, -1)
    forecasted_prices = []
    for _ in future_dates:
        future_price = reg.predict(last_known_data)
        forecasted_prices.append(future_price[0])
        last_known_data = np.roll(last_known_data, -1)
        last_known_data[0, -1] = future_price
    future_df[compound_name] = forecasted_prices
    
    # Print forecasted prices
    print(f"Forecasted prices for {compound_name} from {future_dates[0]} to {future_dates[-1]}:")
    print(future_df)
    
    # Plot historical and forecasted prices
    plt.figure(figsize=(14, 7))
    plt.plot(df.index, df[compound_name], label='Historical Prices', color='blue')
    plt.plot(future_df.index, forecasted_prices, label='Forecasted Prices', color='red', linestyle='--')
    plt.title(f'Historical and Forecasted Prices for {compound_name}')
    plt.xlabel('Date')
    plt.ylabel('Price')
    plt.legend()
    plt.show()
    
    # Calculate and print error metrics
    y_pred = reg.predict(X_test)
    mse = mean_squared_error(y_test, y_pred)
    mae = mean_absolute_error(y_test, y_pred)
    rmse = np.sqrt(mse)
    mape = np.mean(np.abs((y_test - y_pred) / y_test)) * 100
    print(f"\nError Metrics for {compound_name}:")
    print(f"Mean Squared Error (MSE): {mse:.2f}")
    print(f"Mean Absolute Error (MAE): {mae:.2f}")
    print(f"Root Mean Squared Error (RMSE): {rmse:.2f}")
    print(f"Mean Absolute Percentage Error (MAPE): {mape:.2f}%")

# Forecast prices for all compounds
for compound in df.columns:
    forecast_prices_xgb(compound, forecast_period=60, n_lags=12)

Forecasted prices for HDPE film from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
              HDPE film
2024-04-30  1274.308105
2024-05-31  1257.547363
2024-06-30  1159.683838
2024-07-31  1062.042480
2024-08-31  1227.966675
2024-09-30  1200.561035
2024-10-31  1145.120850
2024-11-30  1136.769409
2024-12-31  1057.913818
2025-01-31  1073.662231
2025-02-28  1257.484619
2025-03-31  1252.285645
2025-04-30  1254.674805
2025-05-31  1235.835571
2025-06-30  1243.593262
2025-07-31  1076.340942
2025-08-31  1244.168579
2025-09-30  1243.324097
2025-10-31  1228.490845
2025-11-30  1136.134521
2025-12-31  1135.400269
2026-01-31  1068.430542
2026-02-28  1243.768066
2026-03-31  1243.482178
2026-04-30  1243.081665
2026-05-31  1243.711304
2026-06-30  1251.007812
2026-07-31  1071.635010
2026-08-31  1246.557739
2026-09-30  1243.324097
2026-10-31  1241.036133
2026-11-30  1061.984863
2026-12-31  1136.507446
2027-01-31  1068.430542
2027-02-28  1243.768066
2027-03-31  1243.324097
2027-04-30  1243.081665
2027-05-31  1251.869019
2027-06-30  1248.218140
2027-07-31  1069.245850
2027-08-31  1246.557739
2027-09-30  1232.658691
2027-10-31  1247.430420
2027-11-30  1062.246826
2027-12-31  1136.507446
2028-01-31  1068.430542
2028-02-29  1243.768066
2028-03-31  1243.324097
2028-04-30  1243.081665
2028-05-31  1251.869019
2028-06-30  1248.218140
2028-07-31  1069.245850
2028-08-31  1246.557739
2028-09-30  1232.658691
2028-10-31  1247.430420
2028-11-30  1062.246826
2028-12-31  1136.507446
2029-01-31  1068.430542
2029-02-28  1243.768066
2029-03-31  1243.324097

Error Metrics for HDPE film:
Mean Squared Error (MSE): 50340.01
Mean Absolute Error (MAE): 179.78
Root Mean Squared Error (RMSE): 224.37
Mean Absolute Percentage Error (MAPE): 11.70%
Forecasted prices for LDPE film from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
              LDPE film
2024-04-30  1078.093750
2024-05-31  1348.411377
2024-06-30  1278.071899
2024-07-31  1039.371216
2024-08-31  1168.216431
2024-09-30  1211.995483
2024-10-31  1119.062866
2024-11-30   919.838501
2024-12-31   970.952820
2025-01-31  1046.305908
2025-02-28  1157.451416
2025-03-31  1154.493286
2025-04-30   956.423218
2025-05-31  1261.229370
2025-06-30  1361.444336
2025-07-31   958.028442
2025-08-31  1091.993164
2025-09-30  1172.017212
2025-10-31  1232.316162
2025-11-30   890.706665
2025-12-31   973.646790
2026-01-31  1015.308655
2026-02-28  1075.219116
2026-03-31  1043.350342
2026-04-30   856.331726
2026-05-31  1179.907959
2026-06-30  1520.993896
2026-07-31   892.863953
2026-08-31   989.539246
2026-09-30  1223.680542
2026-10-31  1348.577393
2026-11-30   779.006470
2026-12-31   925.091858
2027-01-31  1053.848999
2027-02-28  1212.895020
2027-03-31   965.129272
2027-04-30   877.630432
2027-05-31  1255.977417
2027-06-30  1654.520508
2027-07-31   804.723633
2027-08-31   834.780457
2027-09-30  1339.663330
2027-10-31  1463.387207
2027-11-30   851.960510
2027-12-31  1041.438599
2028-01-31  1286.446167
2028-02-29  1371.476562
2028-03-31   897.955566
2028-04-30   848.490845
2028-05-31  1311.043213
2028-06-30  1882.435059
2028-07-31   805.058105
2028-08-31   859.049194
2028-09-30  1458.811157
2028-10-31  1654.520508
2028-11-30   856.103821
2028-12-31   956.796753
2029-01-31  1348.409912
2029-02-28  1430.491821
2029-03-31   811.867493

Error Metrics for LDPE film:
Mean Squared Error (MSE): 44288.87
Mean Absolute Error (MAE): 162.04
Root Mean Squared Error (RMSE): 210.45
Mean Absolute Percentage Error (MAPE): 9.79%
Forecasted prices for PET from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
                   PET
2024-04-30  394.919067
2024-05-31  260.009094
2024-06-30  259.782837
2024-07-31  200.710495
2024-08-31  319.145416
2024-09-30  280.581635
2024-10-31  280.553162
2024-11-30  318.333008
2024-12-31  300.987091
2025-01-31  320.798431
2025-02-28  304.942444
2025-03-31  273.162323
2025-04-30  452.530090
2025-05-31  320.018188
2025-06-30  336.034271
2025-07-31  212.056885
2025-08-31  341.918396
2025-09-30  306.356323
2025-10-31  326.383087
2025-11-30  311.434326
2025-12-31  293.596893
2026-01-31  324.184906
2026-02-28  336.320068
2026-03-31  333.222229
2026-04-30  411.486053
2026-05-31  326.508942
2026-06-30  357.096710
2026-07-31  203.946869
2026-08-31  330.402130
2026-09-30  283.143555
2026-10-31  324.887512
2026-11-30  292.124603
2026-12-31  283.562256
2027-01-31  345.798920
2027-02-28  348.002411
2027-03-31  311.338257
2027-04-30  428.450348
2027-05-31  331.076355
2027-06-30  328.864624
2027-07-31  207.073227
2027-08-31  324.225616
2027-09-30  290.266602
2027-10-31  328.796021
2027-11-30  328.055206
2027-12-31  289.267029
2028-01-31  345.798920
2028-02-29  348.002411
2028-03-31  317.968170
2028-04-30  428.450348
2028-05-31  326.082184
2028-06-30  326.385895
2028-07-31  206.302704
2028-08-31  318.637787
2028-09-30  292.009430
2028-10-31  327.939880
2028-11-30  324.416412
2028-12-31  289.267029
2029-01-31  345.798920
2029-02-28  348.002411
2029-03-31  317.968170

Error Metrics for PET:
Mean Squared Error (MSE): 13713.14
Mean Absolute Error (MAE): 86.59
Root Mean Squared Error (RMSE): 117.10
Mean Absolute Percentage Error (MAPE): 271.10%
Forecasted prices for PP homo-polymer fiber from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
            PP homo-polymer fiber
2024-04-30            1246.212036
2024-05-31            1256.778198
2024-06-30            1207.233032
2024-07-31            1054.635010
2024-08-31            1116.397827
2024-09-30            1125.024048
2024-10-31            1045.584595
2024-11-30            1050.535400
2024-12-31            1047.493408
2025-01-31            1059.346191
2025-02-28            1264.596924
2025-03-31            1234.713501
2025-04-30            1241.422241
2025-05-31            1284.923706
2025-06-30            1277.294067
2025-07-31            1103.672241
2025-08-31            1199.801270
2025-09-30            1138.464844
2025-10-31            1039.770264
2025-11-30            1044.172974
2025-12-31            1051.259521
2026-01-31            1053.735962
2026-02-28            1240.220947
2026-03-31            1234.890625
2026-04-30            1226.979858
2026-05-31            1264.312622
2026-06-30            1278.192505
2026-07-31            1106.448120
2026-08-31            1278.479370
2026-09-30            1139.945068
2026-10-31            1060.390747
2026-11-30            1149.999878
2026-12-31            1048.066284
2027-01-31            1046.086792
2027-02-28            1239.591553
2027-03-31            1234.093262
2027-04-30            1226.979858
2027-05-31            1257.635498
2027-06-30            1239.384399
2027-07-31            1113.243286
2027-08-31            1261.820923
2027-09-30            1041.625610
2027-10-31            1059.243164
2027-11-30            1124.607910
2027-12-31            1049.572876
2028-01-31            1046.086792
2028-02-29            1239.591553
2028-03-31            1240.540771
2028-04-30            1265.788208
2028-05-31            1271.451172
2028-06-30            1253.197144
2028-07-31            1184.845337
2028-08-31            1278.719116
2028-09-30            1036.499634
2028-10-31            1061.955933
2028-11-30            1146.052856
2028-12-31            1048.066284
2029-01-31            1039.542603
2029-02-28            1239.591553
2029-03-31            1220.193604

Error Metrics for PP homo-polymer fiber:
Mean Squared Error (MSE): 35435.73
Mean Absolute Error (MAE): 130.66
Root Mean Squared Error (RMSE): 188.24
Mean Absolute Percentage Error (MAPE): 8.48%
Forecasted prices for rPET from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
                  rPET
2024-04-30  408.189484
2024-05-31  424.979675
2024-06-30  383.967651
2024-07-31  424.634369
2024-08-31  463.083710
2024-09-30  572.250366
2024-10-31  463.557007
2024-11-30  430.959656
2024-12-31  462.088898
2025-01-31  502.290192
2025-02-28  456.639954
2025-03-31  407.817474
2025-04-30  425.076111
2025-05-31  460.838562
2025-06-30  378.909973
2025-07-31  447.736328
2025-08-31  531.969788
2025-09-30  385.076752
2025-10-31  415.778168
2025-11-30  424.958618
2025-12-31  554.195923
2026-01-31  416.068451
2026-02-28  390.431549
2026-03-31  457.154541
2026-04-30  404.248962
2026-05-31  399.763794
2026-06-30  418.839111
2026-07-31  489.917572
2026-08-31  387.429260
2026-09-30  382.204437
2026-10-31  397.621307
2026-11-30  479.087738
2026-12-31  394.409515
2027-01-31  411.916443
2027-02-28  426.078796
2027-03-31  430.738983
2027-04-30  397.730499
2027-05-31  430.052246
2027-06-30  420.728729
2027-07-31  415.816315
2027-08-31  377.760468
2027-09-30  401.701904
2027-10-31  388.236176
2027-11-30  416.383636
2027-12-31  389.059509
2028-01-31  390.115082
2028-02-29  422.077881
2028-03-31  419.521820
2028-04-30  388.738098
2028-05-31  418.610413
2028-06-30  414.832336
2028-07-31  398.614868
2028-08-31  383.951172
2028-09-30  389.059509
2028-10-31  378.644836
2028-11-30  398.614868
2028-12-31  378.644836
2029-01-31  389.940613
2029-02-28  430.249329
2029-03-31  399.434937

Error Metrics for rPET:
Mean Squared Error (MSE): 12655.87
Mean Absolute Error (MAE): 90.74
Root Mean Squared Error (RMSE): 112.50
Mean Absolute Percentage Error (MAPE): 18.56%
Forecasted prices for rLDPE Film from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
            rLDPE Film
2024-04-30  470.068146
2024-05-31  549.340515
2024-06-30  559.600281
2024-07-31  423.662994
2024-08-31  419.414398
2024-09-30  353.122559
2024-10-31  368.714722
2024-11-30  446.427521
2024-12-31  389.662292
2025-01-31  575.871277
2025-02-28  433.730835
2025-03-31  383.888519
2025-04-30  553.554810
2025-05-31  555.770569
2025-06-30  558.854248
2025-07-31  434.091248
2025-08-31  451.422180
2025-09-30  378.160645
2025-10-31  426.937592
2025-11-30  495.333099
2025-12-31  409.702515
2026-01-31  514.240295
2026-02-28  431.588776
2026-03-31  360.356079
2026-04-30  577.575623
2026-05-31  538.585999
2026-06-30  573.076050
2026-07-31  449.762421
2026-08-31  470.834564
2026-09-30  411.789673
2026-10-31  413.988953
2026-11-30  535.303040
2026-12-31  407.918335
2027-01-31  506.507843
2027-02-28  436.478973
2027-03-31  365.040161
2027-04-30  582.241089
2027-05-31  498.139435
2027-06-30  592.846985
2027-07-31  427.733002
2027-08-31  502.375977
2027-09-30  510.194183
2027-10-31  430.687134
2027-11-30  563.800171
2027-12-31  392.815613
2028-01-31  530.610107
2028-02-29  450.016205
2028-03-31  396.615356
2028-04-30  538.822571
2028-05-31  476.591370
2028-06-30  563.699524
2028-07-31  460.243103
2028-08-31  478.907837
2028-09-30  546.929504
2028-10-31  455.772125
2028-11-30  545.209351
2028-12-31  421.172058
2029-01-31  500.519623
2029-02-28  446.972382
2029-03-31  377.376862

Error Metrics for rLDPE Film:
Mean Squared Error (MSE): 14121.25
Mean Absolute Error (MAE): 91.11
Root Mean Squared Error (RMSE): 118.83
Mean Absolute Percentage Error (MAPE): 17.29%
Forecasted prices for rPP from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
                   rPP
2024-04-30  724.093689
2024-05-31  660.025635
2024-06-30  641.812683
2024-07-31  614.689148
2024-08-31  733.644714
2024-09-30  799.050842
2024-10-31  724.746033
2024-11-30  808.271423
2024-12-31  721.389648
2025-01-31  801.367615
2025-02-28  617.885925
2025-03-31  623.163940
2025-04-30  656.462830
2025-05-31  772.062622
2025-06-30  733.428650
2025-07-31  810.632935
2025-08-31  756.861572
2025-09-30  771.984741
2025-10-31  735.249207
2025-11-30  661.367493
2025-12-31  681.590637
2026-01-31  662.774414
2026-02-28  752.620117
2026-03-31  785.523010
2026-04-30  738.140076
2026-05-31  829.554504
2026-06-30  738.040955
2026-07-31  816.567444
2026-08-31  630.811523
2026-09-30  759.985901
2026-10-31  639.733765
2026-11-30  755.180908
2026-12-31  781.563660
2027-01-31  740.914429
2027-02-28  795.375549
2027-03-31  721.389648
2027-04-30  760.483948
2027-05-31  666.192993
2027-06-30  761.679565
2027-07-31  649.377991
2027-08-31  774.456421
2027-09-30  745.096741
2027-10-31  775.087646
2027-11-30  773.451538
2027-12-31  736.280884
2028-01-31  786.980652
2028-02-29  633.853760
2028-03-31  770.379333
2028-04-30  628.709290
2028-05-31  765.861877
2028-06-30  766.863464
2028-07-31  740.391846
2028-08-31  809.957764
2028-09-30  738.040955
2028-10-31  760.074524
2028-11-30  649.084534
2028-12-31  755.499390
2029-01-31  655.912292
2029-02-28  785.766541
2029-03-31  768.802917

Error Metrics for rPP:
Mean Squared Error (MSE): 7794.24
Mean Absolute Error (MAE): 69.97
Root Mean Squared Error (RMSE): 88.28
Mean Absolute Percentage Error (MAPE): 10.12%
Forecasted prices for rHDPE Film from 2024-04-30 00:00:00 to 2029-03-31 00:00:00:
            rHDPE Film
2024-04-30  520.186462
2024-05-31  569.058960
2024-06-30  587.328430
2024-07-31  614.878235
2024-08-31  693.768677
2024-09-30  678.473145
2024-10-31  652.605225
2024-11-30  739.518188
2024-12-31  615.961975
2025-01-31  645.812683
2025-02-28  696.726257
2025-03-31  622.827393
2025-04-30  535.507202
2025-05-31  546.903259
2025-06-30  597.828186
2025-07-31  760.422485
2025-08-31  689.126099
2025-09-30  653.006653
2025-10-31  721.911133
2025-11-30  602.541626
2025-12-31  613.896912
2026-01-31  691.815491
2026-02-28  625.484863
2026-03-31  680.123962
2026-04-30  544.516113
2026-05-31  536.696228
2026-06-30  747.423096
2026-07-31  691.223389
2026-08-31  635.480530
2026-09-30  737.119019
2026-10-31  592.639282
2026-11-30  645.097534
2026-12-31  688.266052
2027-01-31  623.034424
2027-02-28  736.430054
2027-03-31  694.024048
2027-04-30  531.376709
2027-05-31  554.813843
2027-06-30  697.521729
2027-07-31  681.055603
2027-08-31  749.497253
2027-09-30  620.178406
2027-10-31  645.004211
2027-11-30  728.547485
2027-12-31  628.203674
2028-01-31  653.750916
2028-02-29  740.986816
2028-03-31  668.617065
2028-04-30  536.904236
2028-05-31  551.495605
2028-06-30  742.716919
2028-07-31  760.433105
2028-08-31  600.710632
2028-09-30  620.325562
2028-10-31  729.050293
2028-11-30  671.199341
2028-12-31  651.836121
2029-01-31  694.716248
2029-02-28  734.789978
2029-03-31  672.869873

Error Metrics for rHDPE Film:
Mean Squared Error (MSE): 2703.86
Mean Absolute Error (MAE): 42.75
Root Mean Squared Error (RMSE): 52.00
Mean Absolute Percentage Error (MAPE): 6.60%


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
from xgboost import plot_tree, plot_importance

# Set the aesthetic parameters in one step
sns.set(style="whitegrid")

# Assuming 'data' is the dictionary containing the historical data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300,
             260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390,
             360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430,
             480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430,
                   430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480,
                   430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180,
                   290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490,
            480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790,
            780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640,
            640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490,
                   490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750,
                   590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640,
                   670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Convert 'Month' column to datetime and set it as index
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)


def create_lag_features(df, compound_name, n_lags):
    lag_df = pd.DataFrame()
    for lag in range(1, n_lags + 1):
        lag_df[f'lag_{lag}'] = df[compound_name].shift(lag)
    return lag_df.dropna()

def train_xgb_model(compound_name, n_lags=12):
    compound_df = df[[compound_name]].copy()
    n_lags = min(n_lags, len(compound_df) - 1)
    lag_df = create_lag_features(compound_df, compound_name, n_lags)
    y = compound_df[compound_name].iloc[n_lags:]
    X_train, X_test, y_train, y_test = train_test_split(lag_df, y, test_size=0.25, random_state=42)

    # Initialize and train the XGBoost model
    reg = xgb.XGBRegressor(n_estimators=1000, learning_rate=0.01, max_depth=3, objective='reg:squarederror', booster='gbtree')
    reg.fit(X_train, y_train, eval_set=[(X_train, y_train), (X_test, y_test)], verbose=False)

    # Plot the first tree
    plt.figure(figsize=(20, 10))
    plot_tree(reg, num_trees=0, rankdir='LR')
    plt.title(f'Decision Tree for {compound_name}')
    plt.show()

    # Plot feature importance
    plt.figure(figsize=(10, 6))
    plot_importance(reg)
    plt.title(f'Feature Importance for {compound_name}')
    plt.show()

    return reg

# Train and visualize the model for all compounds
for compound in df.columns:
    train_xgb_model(compound, n_lags=12)

<Figure size 2000x1000 with 0 Axes>

<Figure size 1000x600 with 0 Axes>

<Figure size 2000x1000 with 0 Axes>

<Figure size 1000x600 with 0 Axes>

<Figure size 2000x1000 with 0 Axes>

<Figure size 1000x600 with 0 Axes>

<Figure size 2000x1000 with 0 Axes>

<Figure size 1000x600 with 0 Axes>

<Figure size 2000x1000 with 0 Axes>

<Figure size 1000x600 with 0 Axes>


pip install plotly

Requirement already satisfied: plotly in c:\users\adity\anaconda3\lib\site-packages (5.9.0)
Requirement already satisfied: tenacity>=6.2.0 in c:\users\adity\anaconda3\lib\site-packages (from plotly) (8.2.2)
Note: you may need to restart the kernel to use updated packages.


import pandas as pd
import plotly.graph_objects as go
from scipy.signal import find_peaks

# Data for HDPE film
hdpe_data = {
    "Month/Year": [
        "April 2024", "May 2024", "June 2024", "July 2024", "August 2024", "September 2024", "October 2024", "November 2024", "December 2024",
        "January 2025", "February 2025", "March 2025", "April 2025", "May 2025", "June 2025", "July 2025", "August 2025", "September 2025", "October 2025", "November 2025", "December 2025",
        "January 2026", "February 2026", "March 2026", "April 2026", "May 2026", "June 2026", "July 2026", "August 2026", "September 2026", "October 2026", "November 2026", "December 2026",
        "January 2027", "February 2027", "March 2027", "April 2027", "May 2027", "June 2027", "July 2027", "August 2027", "September 2027", "October 2027", "November 2027", "December 2027",
        "January 2028", "February 2028", "March 2028", "April 2028", "May 2028", "June 2028", "July 2028", "August 2028", "September 2028", "October 2028", "November 2028", "December 2028",
        "January 2029", "February 2029", "March 2029"
    ],
        "HDPE film":[
        1280.00, 1280.05, 1280.11, 1280.21, 1280.25, 1280.46, 1280.45, 1280.31, 1280.63,
        1280.60, 1280.88, 1280.85, 1280.97, 1281.03, 1281.24, 1281.32, 1281.78, 1281.76, 1281.60, 1281.66, 1281.88,
        1281.99, 1281.85, 1282.05, 1281.95, 1281.65, 1281.31, 1281.32, 1281.52, 1281.43, 1281.19, 1280.97, 1281.09,
        1281.04, 1281.19, 1281.04, 1280.96, 1281.04, 1281.10, 1281.06, 1281.07, 1281.36, 1281.72, 1281.45, 1281.51,
        1281.35, 1281.08, 1281.32, 1281.15, 1281.07, 1281.29, 1281.00, 1280.72, 1280.93, 1281.50, 1281.56, 1281.52,
        1281.42, 1281.71, 1281.91
    ],

}

# Data for LDPE film
ldpe_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "LDPE film":[
        1060.00, 1060.25, 1060.44, 1060.29, 1060.07, 1059.84, 1059.82, 1060.14, 1060.18,
        1060.40, 1060.71, 1060.70, 1060.72, 1060.75, 1061.26, 1061.46, 1061.38, 1061.52, 1061.69, 1061.93, 1062.06,
        1061.98, 1061.60, 1061.99, 1062.07, 1062.10, 1062.14, 1062.26, 1062.29, 1062.10, 1061.88, 1061.73, 1061.96,
        1062.03, 1062.09, 1061.73, 1061.98, 1061.62, 1061.43, 1061.31, 1061.45, 1061.70, 1061.53, 1061.53, 1061.69,
        1062.11, 1062.14, 1062.00, 1062.22, 1062.11, 1062.02, 1062.17, 1062.18, 1062.05, 1061.77, 1062.01, 1061.85,
        1061.61, 1061.08, 1060.91
    ],
}

# Data for PET
pet_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "PET":[
        400.00, 399.95, 399.94, 399.93, 399.98, 399.97, 400.01, 399.97, 399.95,
        400.00, 399.99, 399.95, 400.05, 400.02, 399.99, 400.00, 399.92, 400.03, 400.05, 400.04, 400.07,
        400.06, 399.99, 399.97, 399.96, 399.93, 400.01, 400.11, 400.12, 400.07, 400.18, 400.34, 400.36,
        400.41, 400.45, 400.56, 400.56, 400.48, 400.39, 400.48, 400.37, 400.42, 400.46, 400.52, 400.48,
        400.40, 400.41, 400.37, 400.42, 400.39, 400.34, 400.36, 400.37, 400.33, 400.35, 400.27, 400.20,
        400.19, 400.11, 399.98
    ],
               
}

# Data for PP homo-polymer fiber
pp_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "PP homo-polymer fiber":[
        1250.00, 1250.26, 1250.58, 1250.66, 1250.94, 1250.94, 1250.90, 1250.97, 1251.14,
        1250.97, 1251.22, 1251.38, 1251.58, 1251.68, 1251.64, 1251.69, 1251.63, 1251.99, 1251.77, 1251.79, 1252.01,
        1251.69, 1251.71, 1251.81, 1251.85, 1252.10, 1251.95, 1251.91, 1251.61, 1251.63, 1252.14, 1252.20, 1252.22,
        1252.06, 1252.15, 1252.10, 1251.97, 1252.08, 1252.05, 1252.16, 1251.72, 1251.65, 1251.68, 1251.33, 1251.20,
        1251.20, 1250.99, 1250.69, 1250.45, 1250.37, 1250.35, 1250.50, 1250.05, 1249.74, 1249.72, 1250.10, 1250.33,
        1250.44, 1250.32, 1250.15
    ],
}

# Create DataFrames
df_hdpe = pd.DataFrame(hdpe_data)
df_ldpe = pd.DataFrame(ldpe_data)
df_pet = pd.DataFrame(pet_data)
df_pp = pd.DataFrame(pp_data)

# Merge DataFrames on "Month/Year"
df = df_hdpe.merge(df_ldpe, on="Month/Year").merge(df_pet, on="Month/Year").merge(df_pp, on="Month/Year")

# Convert "Month/Year" to datetime
df["Month/Year"] = pd.to_datetime(df["Month/Year"], format="%B %Y")

# Function to plot time series with peaks and troughs
def plot_time_series_with_peaks(df, column, title):
    fig = go.Figure()

    # Add time series line
    fig.add_trace(go.Scatter(x=df["Month/Year"], y=df[column], mode='lines', name=column))

    # Find peaks and troughs
    peaks, _ = find_peaks(df[column])
    troughs, _ = find_peaks(-df[column])

    # Add peaks
    fig.add_trace(go.Scatter(
        x=df["Month/Year"].iloc[peaks],
        y=df[column].iloc[peaks],
        mode='markers',
        marker=dict(color='red', size=8),
        name='Peaks'
    ))

    # Add troughs
    fig.add_trace(go.Scatter(
        x=df["Month/Year"].iloc[troughs],
        y=df[column].iloc[troughs],
        mode='markers',
        marker=dict(color='blue', size=8),
        name='Troughs'
    ))

    # Update layout
    fig.update_layout(
        title=title,
        xaxis_title='Date',
        yaxis_title='Price (€)',
        showlegend=True
    )

    fig.show()

# Plot for each polymer compound
plot_time_series_with_peaks(df, "HDPE film", "Expected Forward HDPE Film Prices")
plot_time_series_with_peaks(df, "LDPE film", "Expected Forward LDPE Film Prices")
plot_time_series_with_peaks(df, "PET", "Expected Forward PET Prices")
plot_time_series_with_peaks(df, "PP homo-polymer fiber", "Expected Forward PP Homo-Polymer Fiber Prices")

df.describe()


import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Set the aesthetic parameters
sns.set(style="whitegrid")

# Define the data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
}

# Create DataFrame from the data dictionary
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Replace zeros with NaN for interpolation
df = df.replace(0, np.nan)

# Interpolate missing values using linear interpolation
df = df.interpolate(method='linear', limit_direction='both')

# Function to calculate historical volatility
def calculate_historical_volatility(prices):
    log_returns = np.log(prices / prices.shift(1))
    return log_returns.std() * np.sqrt(12)  # Annualize monthly data

# Calculate historical volatilities for each polymer compound
volatilities = pd.DataFrame()
for material in df.columns:
    volatilities[material] = df[material].rolling(window=len(df), min_periods=1).apply(calculate_historical_volatility)

# Plot the historical volatilities
plt.figure(figsize=(12, 8))
for material in volatilities.columns:
    plt.plot(volatilities.index, volatilities[material], label=material)

plt.xlabel('Date')
plt.ylabel('Historical Volatility (Annualized)')
plt.title('Historical Volatilities of Polymer Compounds')
plt.legend()
plt.grid(True)
plt.tight_layout()
plt.show()

# Print the latest volatility values
print("Latest Historical Volatilities:")
print(volatilities.iloc[-1])

Latest Historical Volatilities:
HDPE film                0.278971
LDPE film                0.337475
PET                      2.386315
PP homo-polymer fiber    0.330219
Name: 2024-03-01 00:00:00, dtype: float64


import pandas as pd
import numpy as np

# Assuming 'data' is the dictionary containing the historical data
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300,
             260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390,
             360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430,
             480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430,
                   430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480,
                   430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180,
                   290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490,
            480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790,
            780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640,
            640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490,
                   490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750,
                   590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640,
                   670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Convert 'Month' column to datetime and set it as index
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Function to calculate and print IQR and quartile ranges
def calculate_iqr_and_quartiles(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        print(f"{column}:")
        print(f"  Q1 (25th percentile): {Q1}")
        print(f"  Q3 (75th percentile): {Q3}")
        print(f"  IQR: {IQR}\n")

# Calculate and print IQR and quartile ranges
calculate_iqr_and_quartiles(df)

HDPE film:
  Q1 (25th percentile): 1024.25
  Q3 (75th percentile): 1418.5
  IQR: 394.25

LDPE film:
  Q1 (25th percentile): 985.75
  Q3 (75th percentile): 1672.5
  IQR: 686.75

PET:
  Q1 (25th percentile): 147.5
  Q3 (75th percentile): 292.5
  IQR: 145.0

PP homo-polymer fiber:
  Q1 (25th percentile): 1017.25
  Q3 (75th percentile): 1484.0
  IQR: 466.75

rPET:
  Q1 (25th percentile): 357.5
  Q3 (75th percentile): 482.5
  IQR: 125.0

rLDPE Film:
  Q1 (25th percentile): 430.0
  Q3 (75th percentile): 530.0
  IQR: 100.0

rPP:
  Q1 (25th percentile): 545.0
  Q3 (75th percentile): 732.5
  IQR: 187.5

rHDPE Film:
  Q1 (25th percentile): 567.5
  Q3 (75th percentile): 642.5
  IQR: 75.0


import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Set the aesthetic parameters for seaborn
sns.set(style="whitegrid")

# Define the Historic data for VAR Backtesting
data = {
    'Month': ['Apr-19', 'May-19', 'Jun-19', 'Jul-19', 'Aug-19', 'Sep-19', 'Oct-19', 'Nov-19', 'Dec-19',
              'Jan-20', 'Feb-20', 'Mar-20', 'Apr-20', 'May-20', 'Jun-20', 'Jul-20', 'Aug-20', 'Sep-20',
              'Oct-20', 'Nov-20', 'Dec-20', 'Jan-21', 'Feb-21', 'Mar-21', 'Apr-21', 'May-21', 'Jun-21',
              'Jul-21', 'Aug-21', 'Sep-21', 'Oct-21', 'Nov-21', 'Dec-21', 'Jan-22', 'Feb-22', 'Mar-22',
              'Apr-22', 'May-22', 'Jun-22', 'Jul-22', 'Aug-22', 'Sep-22', 'Oct-22', 'Nov-22', 'Dec-22',
              'Jan-23', 'Feb-23', 'Mar-23', 'Apr-23', 'May-23', 'Jun-23', 'Jul-23', 'Aug-23', 'Sep-23',
              'Oct-23', 'Nov-23', 'Dec-23', 'Jan-24', 'Feb-24', 'Mar-24'],
    'HDPE film': [1166, 1182, 1166, 1102, 1074, 1027, 1037, 988, 967, 958, 958, 819, 817, 757, 809, 878,
                  914, 911, 882, 948, 958, 1140, 1435, 1744, 1858, 1523, 1465, 1350, 1354, 1413, 1388,
                  1595, 1553, 1543, 1547, 1762, 1968, 1729, 1708, 1577, 1443, 1355, 1388, 1393, 1339,
                  1313, 1299, 1334, 1229, 1122, 1065, 1016, 1118, 1158, 1175, 1076, 1063, 1186, 1265, 1280],
    'LDPE film': [1096, 1120, 1105, 1047, 1029, 964, 985, 953, 952, 944, 953, 836, 828, 762, 840, 917,
                  935, 918, 912, 1024, 1051, 1288, 1562, 1920, 2082, 1900, 1816, 1710, 1748, 1767, 1486,
                  1914, 1872, 1851, 1845, 2033, 2225, 1932, 1893, 1660, 1547, 1407, 1486, 1431, 1394,
                  1352, 1330, 1349, 1261, 1095, 979, 986, 1090, 1162, 1221, 1067, 1059, 1290, 1351, 1060],
    'PET': [170, 200, 200, 180, 200, 160, 190, 150, 210, 130, 170, 230, 220, 210, 140, 220, 160, 90, 70,
            180, 300, 160, 100, 80, 140, 70, 70, 200, 160, 180, 290, 90, 70, 0, 360, 200, 140, 250, 10,
            90, 360, 450, 400, 400, 410, 470, 340, 220, 230, 310, 300, 310, 300, 250, 320, 190, 200, 260, 260, 400],
    'PP homo-polymer fiber': [1203, 1223, 1197, 1070, 1055, 1008, 1019, 979, 995, 968, 977, 917, 917,
                              847, 893, 938, 949, 936, 921, 997, 1012, 1194, 1434, 1844, 2066, 1962,
                              1630, 1538, 1631, 1631, 1275, 1712, 1697, 1676, 1654, 1860, 2043, 1790,
                              1648, 1466, 1248, 1191, 1275, 1265, 1220, 1227, 1250, 1275, 1221, 1121,
                              1062, 1019, 1065, 1141, 1145, 1057, 1029, 1237, 1254, 1250],
    'rPET': [360, 360, 350, 330, 320, 340, 340, 340, 400, 370, 330, 310, 380, 370, 360, 340, 290, 300, 260, 300, 340, 380, 380, 380, 380, 390, 430, 480, 440, 420, 520, 520, 480, 720, 350, 390, 360, 600, 820, 720, 560, 480, 490, 550, 700, 610, 700, 550, 550, 570, 480, 450, 460, 430, 480, 450, 410, 380, 420, 410],
    'rLDPE Film': [550, 580, 520, 540, 500, 480, 510, 490, 510, 480, 470, 530, 530, 520, 500, 500, 430, 430, 470, 470, 450, 510, 470, 490, 530, 490, 470, 450, 420, 380, 340, 430, 470, 480, 430, 650, 640, 640, 790, 610, 500, 560, 600, 520, 490, 480, 470, 390, 410, 730, 180, 290, 180, 180, 290, 360, 410, 560, 550, 470],
    'rPP': [590, 600, 600, 580, 570, 570, 550, 570, 570, 560, 530, 520, 510, 530, 520, 460, 450, 490, 480, 480, 510, 510, 520, 530, 550, 600, 620, 560, 570, 620, 650, 730, 630, 650, 530, 790, 780, 670, 860, 830, 750, 790, 900, 840, 770, 780, 760, 740, 750, 740, 740, 700, 650, 640, 640, 600, 580, 600, 620, 730],
    'rHDPE Film': [570, 580, 600, 620, 600, 560, 570, 600, 600, 570, 550, 520, 520, 510, 490, 510, 490, 490, 490, 490, 490, 500, 520, 580, 620, 620, 640, 600, 630, 620, 570, 580, 600, 750, 590, 710, 630, 740, 1010, 1020, 780, 770, 760, 720, 650, 760, 750, 730, 660, 640, 670, 630, 610, 610, 640, 580, 580, 580, 570, 520]
}

# Create DataFrame from the data dictionary
df = pd.DataFrame(data)
df['Month'] = pd.to_datetime(df['Month'], format='%b-%y')
df.set_index('Month', inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Monte Carlo Simulation function using Geometric Brownian Motion (GBM)
def monte_carlo_simulation(s0, mu, sigma, n, num_scenarios, dt):
    prices = np.zeros((num_scenarios, n))
    for i in range(num_scenarios):
        price = np.zeros(n)
        price[0] = s0
        for t in range(1, n):
            simulated_return = (mu - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * np.random.normal()
            next_price = price[t - 1] * np.exp(simulated_return)
            if next_price <= 0:
                prices[i, t:] = np.nan
                break
            price[t] = next_price
        prices[i, :t+1] = price[:t+1]
    return prices

# Simulation parameters
num_scenarios = 10000
n_months_future = 60
dt = 1 / 252

# Initialize simulated prices dictionary
simulated_prices_future = {}

# Loop over each polymer compound and perform simulation
for material in df.columns:
    s0_future = df[material].iloc[-1]
    mu = 0.003
    sigma = 0.27124
    simulated_prices_future[material] = monte_carlo_simulation(s0_future, mu, sigma, n_months_future, num_scenarios, dt)

# Function to calculate VaR and ETL
def calculate_var_etl(prices, confidence_level=0.95):
    sorted_prices = np.sort(prices)
    index_at_var = int((1 - confidence_level) * len(sorted_prices))
    var = sorted_prices[index_at_var]
    etl = sorted_prices[:index_at_var].mean()
    return var, etl

# Function to backtest VaR
def backtest_var(df, var, confidence_level=0.95):
    returns = df.pct_change().dropna()
    var_violations = (returns < -var).sum()
    expected_violations = (1 - confidence_level) * len(returns)
    return var_violations, expected_violations

# Calculate historical VaR for each material
historical_var = {}
for material in df.columns:
    returns = df[material].pct_change().dropna()
    historical_var[material] = calculate_var_etl(returns, confidence_level=0.95)[0]

# Backtest VaR for each material
backtest_results = {}
for material in df.columns:
    var_violations, expected_violations = backtest_var(df[material], historical_var[material])
    backtest_results[material] = (var_violations, expected_violations)

# Print backtest results
for material, result in backtest_results.items():
    print(f"\n{material}:")
    print(f"VaR Violations: {result[0]}")
    print(f"Expected Violations: {result[1]}")

# Plotting and calculating metrics for each material
for material in df.columns:
    prices = simulated_prices_future[material][:, -1]
    median_price = np.median(prices)
    var, etl = calculate_var_etl(prices)
    confidence_interval = np.percentile(prices, [2.5, 97.5])

    print(f"\n{material}:")
    print(f"Median Price: €{median_price:.2f}")
    print(f"95% Confidence Interval: €{confidence_interval[0]:.2f} - €{confidence_interval[1]:.2f}")
    print(f"Value at Risk (VaR) at 95% confidence: €{var:.2f}")
    print(f"Expected Tail Loss (ETL): €{etl:.2f}")

    plt.figure(figsize=(10, 6))
    sns.histplot(prices, bins=50, kde=True)
    plt.axvline(median_price, color='r', linestyle='dashed', linewidth=2, label='Median')
    plt.axvline(var, color='g', linestyle='dashed', linewidth=2, label='VaR (95%)')
    plt.axvline(etl, color='b', linestyle='dashed', linewidth=2, label='ETL')
    plt.title(f'Distribution of Simulated Prices for {material}')
    plt.xlabel('Price')
    plt.ylabel('Frequency')
    plt.legend()
    plt.show()

HDPE film:
VaR Violations: 54
Expected Violations: 2.950000000000003

LDPE film:
VaR Violations: 54
Expected Violations: 2.950000000000003

PET:
VaR Violations: 52
Expected Violations: 2.950000000000003

PP homo-polymer fiber:
VaR Violations: 54
Expected Violations: 2.950000000000003

rPET:
VaR Violations: 53
Expected Violations: 2.950000000000003

rLDPE Film:
VaR Violations: 54
Expected Violations: 2.950000000000003

rPP:
VaR Violations: 55
Expected Violations: 2.950000000000003

rHDPE Film:
VaR Violations: 54
Expected Violations: 2.950000000000003

HDPE film:
Median Price: €1271.12
95% Confidence Interval: €982.89 - €1645.16
Value at Risk (VaR) at 95% confidence: €1022.32
Expected Tail Loss (ETL): €969.16

LDPE film:
Median Price: €1053.16
95% Confidence Interval: €811.40 - €1357.67
Value at Risk (VaR) at 95% confidence: €845.82
Expected Tail Loss (ETL): €802.02

PET:
Median Price: €396.91
95% Confidence Interval: €305.69 - €513.07
Value at Risk (VaR) at 95% confidence: €318.67
Expected Tail Loss (ETL): €301.84

PP homo-polymer fiber:
Median Price: €1241.22
95% Confidence Interval: €955.18 - €1598.48
Value at Risk (VaR) at 95% confidence: €998.64
Expected Tail Loss (ETL): €945.51

rPET:
Median Price: €407.40
95% Confidence Interval: €313.48 - €526.97
Value at Risk (VaR) at 95% confidence: €328.58
Expected Tail Loss (ETL): €310.51

rLDPE Film:
Median Price: €465.96
95% Confidence Interval: €361.00 - €605.23
Value at Risk (VaR) at 95% confidence: €375.94
Expected Tail Loss (ETL): €355.77

rPP:
Median Price: €724.69
95% Confidence Interval: €556.25 - €932.83
Value at Risk (VaR) at 95% confidence: €580.16
Expected Tail Loss (ETL): €549.18

rHDPE Film:
Median Price: €515.04
95% Confidence Interval: €398.94 - €666.67
Value at Risk (VaR) at 95% confidence: €416.33
Expected Tail Loss (ETL): €394.03


import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

# Data for Forecasted Prices for backtesting

# Set the aesthetic parameters for seaborn
sns.set(style="whitegrid")

# Data for HDPE film
hdpe_data = {
    "Month/Year": [
        "April 2024", "May 2024", "June 2024", "July 2024", "August 2024", "September 2024", "October 2024", "November 2024", "December 2024",
        "January 2025", "February 2025", "March 2025", "April 2025", "May 2025", "June 2025", "July 2025", "August 2025", "September 2025", "October 2025", "November 2025", "December 2025",
        "January 2026", "February 2026", "March 2026", "April 2026", "May 2026", "June 2026", "July 2026", "August 2026", "September 2026", "October 2026", "November 2026", "December 2026",
        "January 2027", "February 2027", "March 2027", "April 2027", "May 2027", "June 2027", "July 2027", "August 2027", "September 2027", "October 2027", "November 2027", "December 2027",
        "January 2028", "February 2028", "March 2028", "April 2028", "May 2028", "June 2028", "July 2028", "August 2028", "September 2028", "October 2028", "November 2028", "December 2028",
        "January 2029", "February 2029", "March 2029"
    ],
    "HDPE film": [
        1280.00, 1280.05, 1280.11, 1280.21, 1280.25, 1280.46, 1280.45, 1280.31, 1280.63,
        1280.60, 1280.88, 1280.85, 1280.97, 1281.03, 1281.24, 1281.32, 1281.78, 1281.76, 1281.60, 1281.66, 1281.88,
        1281.99, 1281.85, 1282.05, 1281.95, 1281.65, 1281.31, 1281.32, 1281.52, 1281.43, 1281.19, 1280.97, 1281.09,
        1281.04, 1281.19, 1281.04, 1280.96, 1281.04, 1281.10, 1281.06, 1281.07, 1281.36, 1281.72, 1281.45, 1281.51,
        1281.35, 1281.08, 1281.32, 1281.15, 1281.07, 1281.29, 1281.00, 1280.72, 1280.93, 1281.50, 1281.56, 1281.52,
        1281.42, 1281.71, 1281.91
    ]
}

# Data for LDPE film
ldpe_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "LDPE film": [
        1060.00, 1060.25, 1060.44, 1060.29, 1060.07, 1059.84, 1059.82, 1060.14, 1060.18,
        1060.40, 1060.71, 1060.70, 1060.72, 1060.75, 1061.26, 1061.46, 1061.38, 1061.52, 1061.69, 1061.93, 1062.06,
        1061.98, 1061.60, 1061.99, 1062.07, 1062.10, 1062.14, 1062.26, 1062.29, 1062.10, 1061.88, 1061.73, 1061.96,
        1062.03, 1062.09, 1061.73, 1061.98, 1061.62, 1061.43, 1061.31, 1061.45, 1061.70, 1061.53, 1061.53, 1061.69,
        1062.11, 1062.14, 1062.00, 1062.22, 1062.11, 1062.02, 1062.17, 1062.18, 1062.05, 1061.77, 1062.01, 1061.85,
        1061.61, 1061.08, 1060.91
    ]
}

# Data for PET
pet_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "PET": [
        400.00, 399.95, 399.94, 399.93, 399.98, 399.97, 400.01, 399.97, 399.95,
        400.00, 399.99, 399.95, 400.05, 400.02, 399.99, 400.00, 399.92, 400.03, 400.05, 400.04, 400.07,
        400.06, 399.99, 399.97, 399.96, 399.93, 400.01, 400.11, 400.12, 400.07, 400.18, 400.34, 400.36,
        400.41, 400.45, 400.56, 400.56, 400.48, 400.39, 400.48, 400.37, 400.42, 400.46, 400.52, 400.48,
        400.40, 400.41, 400.37, 400.42, 400.39, 400.34, 400.36, 400.37, 400.33, 400.35, 400.27, 400.20,
        400.19, 400.11, 399.98
    ]
}

# Data for PP homo-polymer fiber
pp_data = {
    "Month/Year": hdpe_data["Month/Year"],
    "PP homo-polymer fiber": [
        1250.00, 1250.26, 1250.58, 1250.66, 1250.94, 1250.94, 1250.90, 1250.97, 1251.14,
        1250.97, 1251.22, 1251.38, 1251.58, 1251.68, 1251.64, 1251.69, 1251.63, 1251.99, 1251.77, 1251.79, 1252.01,
        1251.69, 1251.71, 1251.81, 1251.85, 1252.10, 1251.95, 1251.91, 1251.61, 1251.63, 1252.14, 1252.20, 1252.22,
        1252.06, 1252.15, 1252.10, 1251.97, 1252.08, 1252.05, 1252.16, 1251.72, 1251.65, 1251.68, 1251.33, 1251.20,
        1251.20, 1250.99, 1250.69, 1250.45, 1250.37, 1250.35, 1250.50, 1250.05, 1249.74, 1249.72, 1250.10, 1250.33,
        1250.44, 1250.32, 1250.15
    ]
}

# Create DataFrames
df_hdpe = pd.DataFrame(hdpe_data)
df_ldpe = pd.DataFrame(ldpe_data)
df_pet = pd.DataFrame(pet_data)
df_pp = pd.DataFrame(pp_data)

# Merge DataFrames on "Month/Year"
df = df_hdpe.merge(df_ldpe, on="Month/Year").merge(df_pet, on="Month/Year").merge(df_pp, on="Month/Year")

# Convert "Month/Year" to datetime
df["Month/Year"] = pd.to_datetime(df["Month/Year"], format="%B %Y")
df.set_index("Month/Year", inplace=True)

# Data Preprocessing: Identify and treat outliers using the IQR method
def treat_outliers_iqr(df):
    for column in df.columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - 1.5 * IQR
        upper_bound = Q3 + 1.5 * IQR
        df[column] = np.where(df[column] < lower_bound, lower_bound, df[column])
        df[column] = np.where(df[column] > upper_bound, upper_bound, df[column])
    return df

# Treat outliers in the DataFrame
df = treat_outliers_iqr(df)

# Monte Carlo Simulation function using Geometric Brownian Motion (GBM)
def monte_carlo_simulation(s0, mu, sigma, n, num_scenarios, dt):
    prices = np.zeros((num_scenarios, n))
    for i in range(num_scenarios):
        price = np.zeros(n)
        price[0] = s0
        for t in range(1, n):
            simulated_return = (mu - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * np.random.normal()
            next_price = price[t - 1] * np.exp(simulated_return)
            if next_price <= 0:
                prices[i, t:] = np.nan
                break
            price[t] = next_price
        prices[i, :t+1] = price[:t+1]
    return prices

# Simulation parameters
num_scenarios = 10000
n_months_future = 60
dt = 1 / 252

# Initialize simulated prices dictionary
simulated_prices_future = {}

# Loop over each polymer compound and perform simulation
for material in df.columns:
    s0_future = df[material].iloc[-1]
    mu = 0.003
    sigma = 0.27124
    simulated_prices_future[material] = monte_carlo_simulation(s0_future, mu, sigma, n_months_future, num_scenarios, dt)

# Function to calculate VaR and ETL
def calculate_var_etl(prices, confidence_level=0.95):
    sorted_prices = np.sort(prices)
    index_at_var = int((1 - confidence_level) * len(sorted_prices))
    var = sorted_prices[index_at_var]
    etl = sorted_prices[:index_at_var].mean()
    return var, etl

# Function to backtest VaR
def backtest_var(df, var, confidence_level=0.95):
    returns = df.pct_change().dropna()
    var_violations = (returns < -var).sum()
    expected_violations = (1 - confidence_level) * len(returns)
    return var_violations, expected_violations

# Calculate historical VaR for each material
historical_var = {}
for material in df.columns:
    returns = df[material].pct_change().dropna()
    historical_var[material] = calculate_var_etl(returns, confidence_level=0.95)[0]

# Backtest VaR for each material
backtest_results = {}
for material in df.columns:
    var_violations, expected_violations = backtest_var(df[material], historical_var[material])
    backtest_results[material] = (var_violations, expected_violations)

# Print backtest results
for material, result in backtest_results.items():
    print(f"\n{material}:")
    print(f"VaR Violations: {result[0]}")
    print(f"Expected Violations: {result[1]}")

# Plotting and calculating metrics for each material
for material in df.columns:
    prices = simulated_prices_future[material][:, -1]
    median_price = np.median(prices)
    var, etl = calculate_var_etl(prices)
    confidence_interval = np.percentile(prices, [2.5, 97.5])

    print(f"\n{material}:")
    print(f"Median Price: €{median_price:.2f}")
    print(f"95% Confidence Interval: €{confidence_interval[0]:.2f} - €{confidence_interval[1]:.2f}")
    print(f"Value at Risk (VaR) at 95% confidence: €{var:.2f}")
    print(f"Expected Tail Loss (ETL): €{etl:.2f}")

    plt.figure(figsize=(10, 6))
    sns.histplot(prices, bins=50, kde=True)
    plt.axvline(median_price, color='r', linestyle='dashed', linewidth=2, label='Median')
    plt.axvline(var, color='g', linestyle='dashed', linewidth=2, label='VaR (95%)')
    plt.axvline(etl, color='b', linestyle='dashed', linewidth=2, label='ETL')
    plt.title(f'Distribution of Simulated Prices for {material}')
    plt.xlabel('Price')
    plt.ylabel('Frequency')
    plt.legend()
    plt.show()

HDPE film:
VaR Violations: 54
Expected Violations: 2.950000000000003

LDPE film:
VaR Violations: 56
Expected Violations: 2.950000000000003

PET:
VaR Violations: 52
Expected Violations: 2.950000000000003

PP homo-polymer fiber:
VaR Violations: 56
Expected Violations: 2.950000000000003

HDPE film:
Median Price: €1275.18
95% Confidence Interval: €986.02 - €1650.74
Value at Risk (VaR) at 95% confidence: €1027.84
Expected Tail Loss (ETL): €973.32

LDPE film:
Median Price: €1051.37
95% Confidence Interval: €817.25 - €1358.66
Value at Risk (VaR) at 95% confidence: €854.35
Expected Tail Loss (ETL): €807.48

PET:
Median Price: €397.03
95% Confidence Interval: €306.88 - €509.70
Value at Risk (VaR) at 95% confidence: €321.60
Expected Tail Loss (ETL): €303.21

PP homo-polymer fiber:
Median Price: €1237.02
95% Confidence Interval: €960.19 - €1599.54
Value at Risk (VaR) at 95% confidence: €998.23
Expected Tail Loss (ETL): €948.36


import pandas as pd

# Data for VaR backtesting results
data = {
    "Polymer Type": ["HDPE film", "LDPE film", "PET", "PP homo-polymer fiber"],
    "VaR Violations (Historical)": [54, 54, 52, 54],
    "Expected Violations (Historical)": [2.95, 2.95, 2.95, 2.95],
    "VaR Violations (Forecasted)": [54, 56, 52, 56],
    "Expected Violations (Forecasted)": [2.95, 2.95, 2.95, 2.95]
}

# Create a DataFrame
df = pd.DataFrame(data)

# Print the DataFrame
print(df)

# Describe the DataFrame
print("\nDescription of VaR Backtesting Results:")
print(df.describe())

            Polymer Type  VaR Violations (Historical)  \
0              HDPE film                           54   
1              LDPE film                           54   
2                    PET                           52   
3  PP homo-polymer fiber                           54   

   Expected Violations (Historical)  VaR Violations (Forecasted)  \
0                              2.95                           54   
1                              2.95                           56   
2                              2.95                           52   
3                              2.95                           56   

   Expected Violations (Forecasted)  
0                              2.95  
1                              2.95  
2                              2.95  
3                              2.95  

Description of VaR Backtesting Results:
       VaR Violations (Historical)  Expected Violations (Historical)  \
count                          4.0                              4.00   
mean                          53.5                              2.95   
std                            1.0                              0.00   
min                           52.0                              2.95   
25%                           53.5                              2.95   
50%                           54.0                              2.95   
75%                           54.0                              2.95   
max                           54.0                              2.95   

       VaR Violations (Forecasted)  Expected Violations (Forecasted)  
count                     4.000000                              4.00  
mean                     54.500000                              2.95  
std                       1.914854                              0.00  
min                      52.000000                              2.95  
25%                      53.500000                              2.95  
50%                      55.000000                              2.95  
75%                      56.000000                              2.95  
max                      56.000000                              2.95

	HDPE film	LDPE film	PET	PP homo-polymer fiber	rPET	rLDPE Film	rPP	rHDPE Film
count	60.000000	60.000000	60.000000	60.000000	60.000000	60.000000	60.000000	60.000000
mean	1243.116667	1326.533333	213.666667	1272.066667	439.666667	479.500000	629.333333	619.333333
std	286.557776	392.274051	107.624058	320.775695	121.054220	112.256139	111.459754	108.797225
min	757.000000	762.000000	0.000000	847.000000	260.000000	180.000000	450.000000	490.000000
25%	1024.250000	985.750000	147.500000	1017.250000	357.500000	430.000000	545.000000	567.500000
50%	1184.000000	1241.000000	200.000000	1211.500000	405.000000	485.000000	600.000000	600.000000
75%	1418.500000	1672.500000	292.500000	1484.000000	482.500000	530.000000	732.500000	642.500000
max	1968.000000	2225.000000	470.000000	2066.000000	820.000000	790.000000	900.000000	1020.000000

	Month/Year	HDPE film	LDPE film	PET	PP homo-polymer fiber
count	60	60.000000	60.000000	60.000000	60.000000
mean	2026-09-15 23:12:00	1281.173833	1061.467167	400.183000	1251.268500
min	2024-04-01 00:00:00	1280.000000	1059.820000	399.920000	1249.720000
25%	2025-06-23 12:00:00	1280.967500	1061.037500	399.990000	1250.640000
50%	2026-09-16 00:00:00	1281.190000	1061.695000	400.110000	1251.595000
75%	2027-12-08 18:00:00	1281.520000	1062.035000	400.375000	1251.865000
max	2029-03-01 00:00:00	1282.050000	1062.290000	400.560000	1252.220000
std	NaN	0.506888	0.722475	0.208036	0.730662

	HDPE film	LDPE film	PET	PP homo-polymer fiber	rPET	rLDPE Film	rPP	rHDPE Film
Month
2019-04-01	1166	1096	170	1203	360	550	590	570
2019-05-01	1182	1120	200	1223	360	580	600	580
2019-06-01	1166	1105	200	1197	350	520	600	600
2019-07-01	1102	1047	180	1070	330	540	580	620
2019-08-01	1074	1029	200	1055	320	500	570	600

	HDPE film	LDPE film	PET	PP homo-polymer fiber	rPET	rLDPE Film	rPP	rHDPE Film
Month
2023-11-01	1076	1067	190	1057	450	360	600	580
2023-12-01	1063	1059	200	1029	410	410	580	580
2024-01-01	1186	1290	260	1237	380	560	600	580
2024-02-01	1265	1351	260	1254	420	550	620	570
2024-03-01	1280	1060	400	1250	410	470	730	520

	HDPE film	LDPE film	PET	PP homo-polymer fiber	rPET	rLDPE Film	rPP	rHDPE Film
Month
2019-04-01	1166	1096	170	1203	360	550	590	570
2019-05-01	1182	1120	200	1223	360	580	600	580
2019-06-01	1166	1105	200	1197	350	520	600	600
2019-07-01	1102	1047	180	1070	330	540	580	620
2019-08-01	1074	1029	200	1055	320	500	570	600
2019-09-01	1027	964	160	1008	340	480	570	560
2019-10-01	1037	985	190	1019	340	510	550	570
2019-11-01	988	953	150	979	340	490	570	600
2019-12-01	967	952	210	995	400	510	570	600
2020-01-01	958	944	130	968	370	480	560	570
2020-02-01	958	953	170	977	330	470	530	550
2020-03-01	819	836	230	917	310	530	520	520
2020-04-01	817	828	220	917	380	530	510	520
2020-05-01	757	762	210	847	370	520	530	510
2020-06-01	809	840	140	893	360	500	520	490
2020-07-01	878	917	220	938	340	500	460	510
2020-08-01	914	935	160	949	290	430	450	490
2020-09-01	911	918	90	936	300	430	490	490
2020-10-01	882	912	70	921	260	470	480	490
2020-11-01	948	1024	180	997	300	470	480	490
2020-12-01	958	1051	300	1012	340	450	510	490
2021-01-01	1140	1288	160	1194	380	510	510	500
2021-02-01	1435	1562	100	1434	380	470	520	520
2021-03-01	1744	1920	80	1844	380	490	530	580
2021-04-01	1858	2082	140	2066	380	530	550	620
2021-05-01	1523	1900	70	1962	390	490	600	620
2021-06-01	1465	1816	70	1630	430	470	620	640
2021-07-01	1350	1710	200	1538	480	450	560	600
2021-08-01	1354	1748	160	1631	440	420	570	630
2021-09-01	1413	1767	180	1631	420	380	620	620
2021-10-01	1388	1486	290	1275	520	340	650	570
2021-11-01	1595	1914	90	1712	520	430	730	580
2021-12-01	1553	1872	70	1697	480	470	630	600
2022-01-01	1543	1851	0	1676	720	480	650	750
2022-02-01	1547	1845	360	1654	350	430	530	590
2022-03-01	1762	2033	200	1860	390	650	790	710
2022-04-01	1968	2225	140	2043	360	640	780	630
2022-05-01	1729	1932	250	1790	600	640	670	740
2022-06-01	1708	1893	10	1648	820	790	860	1010